tests/test.R

In lpcde: Boundary Adaptive Local Polynomial Conditional Density Estimator

library("lpcde")

set.seed(42)
n <- 1000
x_data <- matrix(rnorm(1 * n, mean = 0, sd = 1), ncol = 1)
y_data <- matrix(rnorm(n, mean = 0, sd = 1))
y_grid <- stats::quantile(y_data, seq(from = 0.1, to = 0.9, by = 0.1))

# bw estimation
model1 <- lpbwcde(x_data = x_data, y_data = y_data, y_grid = y_grid, x = 0, bw_type = "imse-rot")
summary(model1)

# density estimation
model2 <- lpcde(x_data = x_data, y_data = y_data, y_grid = y_grid, x = 0, bw = model1$BW[, 2])
summary(model2)

# non-negative and integrating to 1 density estimation
model2 <- lpcde(x_data = x_data, y_data = y_data, y_grid = y_grid, x = 0, bw = model1$BW[, 2], nonneg = TRUE, normalize = TRUE)
summary(model2)

# bw estimation
model1 <- lpbwcde(x_data = x_data, y_data = y_data, y_grid = y_grid, x = 0, bw_type = "mse-rot")
summary(model1)

# density estimation
model2 <- lpcde(x_data = x_data, y_data = y_data, y_grid = y_grid, x = 0, bw = model1$BW[, 2])
summary(model2)

set.seed(42)
n <- 1000
x_data <- matrix(rnorm(2 * n, mean = 0, sd = 1), ncol = 2)
y_data <- matrix(rnorm(n, mean = 0, sd = 1))
y_grid <- stats::quantile(y_data, seq(from = 0.1, to = 0.9, by = 0.1))

# bw estimation
model1 <- lpbwcde(x_data = x_data, y_data = y_data, y_grid = y_grid, x = matrix(c(0, 0), ncol = 2), bw_type = "imse-rot")
summary(model1)

# density estimation
model2 <- lpcde(x_data = x_data, y_data = y_data, y_grid = y_grid, x = matrix(c(0, 0), ncol = 2), bw = model1$BW[, 2])
summary(model2)

# bw estimation
model1 <- lpbwcde(x_data = x_data, y_data = y_data, y_grid = y_grid, x = matrix(c(0, 0), ncol = 2), bw_type = "mse-rot")
summary(model1)

# density estimation
model2 <- lpcde(x_data = x_data, y_data = y_data, y_grid = y_grid, x = matrix(c(0, 0), ncol = 2), bw = model1$BW[, 2])
summary(model2)